ceph/ceph-nvmeof

Ceph-nvmeof is a service that exports Ceph block storage over the NVMe-oF/TCP protocol. The project bridges the gap for platforms like VMware that lack native Ceph support by allowing existing RBD images to be exposed as NVMe-oF namespaces, replacing older iSCSI-based approaches with a more modern standard.

The service solves the problem of accessing Ceph storage from systems without RBD drivers by translating between Ceph's native block storage interface and the NVMe-oF protocol. It operates as a gateway that sits between Ceph clusters and NVMe initiators, handling the protocol translation and namespace management. Users configure subsystems, add namespaces backed by RBD images, create listeners for client connections, and define access controls through a command-line interface.

Adoption suits environments where virtualization platforms or other systems need block storage from Ceph but cannot run native RBD clients. The project provides both automated setup through containerized deployment and manual configuration steps for production environments. The README emphasizes that the containerized approach is intended for testing and development rather than production use, and that RBD image creation and management remain outside this component's scope.

Development activity shows consistent engagement with the codebase through regular commits addressing both features and maintenance. The project maintains active code review practices with pull requests receiving feedback and iteration. Testing infrastructure is in place with verification tools like flake8 integrated into the build process. Documentation is kept current with setup instructions, usage examples, and deployment guidance updated to reflect the project's state.